TEST #5 RESULTS
03/19/2025 - 10:07 PM
ACCELERATOR
MODEL
14.3
tokens/s
generation
5.58
sec
time to first token
266
tokens/s
prompt
88
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 7352 24-Core Processor (znver2)
RAM
125.64GB
OS
Linux
Kernel Release
6.8.0-52-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #53~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Wed Jan 15 19:18:46 UTC 2
RUNTIME
Name
llamafile
Version
0.9.1
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
312
tokens/s
14.8
tokens/s
3.35
sec
pp4096+tg256
199
tokens/s
12.4
tokens/s
20.62
sec
pp2048+tg256
253
tokens/s
14.0
tokens/s
8.17
sec
pp2048+tg768
245
tokens/s
13.9
tokens/s
8.44
sec
pp1024+tg1024
308
tokens/s
14.4
tokens/s
3.39
sec
pp1280+tg3072
294
tokens/s
13.4
tokens/s
4.42
sec
pp384+tg1152
290
tokens/s
15.0
tokens/s
1.39
sec
pp64+tg1024
319
tokens/s
15.4
tokens/s
261
ms
pp16+tg1536
172
tokens/s
15.2
tokens/s
153
ms